Marathi as a Formal Language and development of
                                                a Marathi Compiler
                                       (i.e. Marathitun Adnyavali Vikasit Karane)
                                                     Prof.  Mukund Dhaygude
                                                Email : mukunddhaygude1@yahoo.co.in
                      The paper deals with the following topics       Basic interpreter/compiler is now available.
                 i.e.                                                 The Mbasic.exe is available on
                 1.  Science, Technology and Management               www.mbasic.8m.com website. One can have
                      aspects of “Sanganak Ani Marathi”               it from us by contacting address given above
                                                                      (by phone or by email).
                 2.  Review of the present scenario                         To familiarize the use of Marathi, we are
                 3.  Mathematical Marathi or Marathi as a             proving Marathi to be mathematically and
                      Formal Language                                 computationally complete. Later we will
                 4.  Computational Marathi and Marathi                develop system softwares in Marathi and then
                      Compiler                                        we will develop application softwares in
                                                                      Marathi. The present system softwares are in
                 5.  Marathi as a Natural Language and                English. Operating systems are now getting
                      Marathi Corpora Creation and                    converted to show Marathi icons. We will then
                      Management                                      develop interpreter/compiler that will enable
                                                                      us to program in Marathi and that is what this
                 1.  The Science, Technology and                      white paper talks about.
                 Management aspects of “Sanganak and                  3.  Mathematical Marathi or Marathi as a
                 Marathi”                                             formal language
                      Sanganak comes from the backgrounds                   An alphabet A of a language L comprises
                 of science, technology and management.               of a set of all possible letters/characters that
                 Science deals with proving concepts                  are fundamentally or derivable in the
                 mathematically / experimentally. Technology          Language. A language L comprises of a
                 takes off from where science leaves by               grammar G. The Grammar G uses the
                 producing a prototype of the concept is more         alphabet A. A Grammar G comprises of set
                 like the sense of ‘The proof of pudding lies in      of rules of the grammar. These rules are
                 having it’. Technology demonstrates the              defined in terms of tokens, symbols,
                 production on a mass scale of the product            operators, precedence, and mathematical
                 which was prototyped after scientists proved         \Boolean\digital operators, statements,
                 it. Management concerns itself with value            computational statements, conditional
                 addition in the society. Management begins           statements, loops, programming constructs
                 where Science and Technology leaves.                 etc. The rules are structured from elementary
                 Science and technology thus mass produce             linguistic formations such as letters/
                 several products. Management introduces              characters, words, sentences, paragraphs,
                 these products in to the society and does a          stories, essays, news, dialogues, pages etc.
                 value addition to the society.
                 2.  Review of the present scenario                   4.  Computational Marathi and Marathi
                                                                      Compiler
                      We hypothesized last time that we                     Having defined the mathematical Marathi
                 should define Marathi as a formal language           we then represent it using the language
                 and write programs in Marathi since a Marathi
                                                        O gßJUH$ Am{U _amR>rO
                 92
                 representation schemes (such as Unicode etc)           importantly, the processor’s attention. MBasic
                 in the keyboard, printers, sanganak, operating         5.0 has more debugging and error
                 system etc and then proceed to see the                 identification support. It checks and double-
                 Mbasic.exe software that demonstrates the              checks your code for errors. Yet, at this time it
                 ability to interpret the simple sample programs        is not able to suggest solutions to the errors.
                 that are distributed with it. Having executed               One may use, distribute and copy MBasic
                 the user programs the control is returned to           as many times as you wish, though you may
                 operating system.                                      not in any way modify MBasic.
                      Marathi Basic or just MBasic for short is              We are desperately in need of someone
                 an experimental version of an interpreter              writing a good help file on MBasic. Please help
                 created to consider the possibility of a full-         me. For now, you are provided with sample
                 fledged programming language based on the              programs. I have tried to provide sample
                 Marathi Language. This language will define            programs that use almost every variant of the
                 its own syntax and statements that will be             MBasic syntax. You should be aware that a
                 extremely close to the spoken Marathi                  person with some programming knowledge
                 language. MBasic being experimental, we                must read the sample files and provide
                 have experimented at converting the entire             explanations for someone with absolutely no
                 interface to Marathi. But rather than forcing          programming knowledge. Some reading,
                 Windows Operating System itself to change              experimentation, interpolation and creativity
                 the system font, which is done by a good many          should be enough to give you the hang of
                 softwares, MBasic only keeps its own interface         MBasic. As the author, I must say that
                 in Marathi while leaving other applications in         programming in MBasic can be quite fun.
                 control of their own. Hence, MBasic respects
                 other applications privacy and will work in co-             Modifications, Undocumented features,
                 operation with accompanying software.                  changes, additions, restrictions and things you
                      The name Marathi Basic was coined from            should know about:
                 a Marathi-to-basic transliterator that we had               1. One may have to enclose a syllable
                 created to help Marathi medium students learn          consisting completely of a vowel in {} as the
                 QBasic. But what resulted were new syntax,             transliterator program is unable to determine
                 statements and entire interfaces in Marathi.           when a syllable has ended. For example, to
                 MBasic has been designed to be extensible              enter the word ‘file’ in devnagri, if one types
                 as a language and new statements should be             only ‘phaaila’, the transliterator will not
                 easily integrated as and when required. This           understand that ‘phaa’ is one syllable while ‘i’
                 is a task for which I require your comments            itself is an independent syllable since no
                 and suggestions. Please send me ideas and              consonant comes between the two. Hence,
                 suggestions on how to improve, increase or             one will have to type in ‘phaa {i} la’ to get the
                 change MBasic so that it can be a better piece         proper transliteration. For more information on
                 of work.                                               these conversion rules, please visit
                      MBasic 5.0 is a complete remake of the            Omkarananda Ashram website
                 entire MBasic engine. Right from the compiler,         www.omkarananda-ashram.org. MBasic uses
                 interpreter, to the display, everything has been       the transliterator developed by them for its
                 internally revamped. MBasic 5.0 is very slow.          conversion routines. For complex devnagri
                 Unlike its predecessors, it provides features          entry, it would be a good approach to visit and
                 which cannot be compiled or even pseudo-               read about the ITrans transliteration scheme
                 compiled. They have to be interpreted. In a            that is used by the translators.
                 windows environment this costs system                       2. Interlaced devnagri and roman text in
                 resources such as memory and more                      the program is allowed. Even strings may
                                                        OgßJUH$ Am{U _amR>rO
                                                                                                                     93
                contain interlaced Roman and Devnagri text.           end of a line. In this case, just add some extra
                Input and display functions support this feature      spaces at the end of the line to correct this
                completely. One may enter Roman as well as            problem. We still have to analyze why this
                Devnagri text mixed together. String                  happens. A possible cause might be that
                comparisons will take this into consideration         MBasic uses dll’s compiled from a lot many
                when comparing two strings. To interlace              languages including VC++, Delphi, VB, etc. It
                roman text within devnagri text, just enclose it      may be the transfer of strings from so many
                in ##. This feature is useful for entering special    dll’s that causes this.
                characters such as ‘{‘ and ‘}’ which are                    8. The garbage collection in MBasic is
                recognized by the transliterator and need to          very bad. It hogs a lot of memory. So it is
                be explicitly placed within ##s in order to show      advisable for one to save one’s work regularly.
                up.                                                   Please save thy work before running any
                      3. Obviously, there is absolutely no way        program. If it gives any problems, just restart
                to add a # in the program. Sorry about this.          it and it will work fine (since MBasic only runs
                Also, there is yet no facility to add a “ within a    on windows, you should be quite familiar with
                string. Sorry again. Though the input function        this practice).
                provides this facility, the editor will not allow           MBasic has been a voluntary effort to
                this. We are working on schemes to provide            develop MKeyboards, MBios, Mcomputers,
                these facilities. Such special features increase      Mprinters, and other Msoftwares so that we
                the load on the CPU exponentially and will be         can check them out and fill up their
                added only in case a lot of people demand it.         weaknesses. We also have our hardware
                      4. Provision for making stand-alone             development projects in Marathi. If you like
                executables is provided. Of course, these             MBasic, please contact us for further
                executables still need support dll’s and also         development.
                the true type fonts. But they can run completely            5 Marathi as a Natural Language and
                independent of the MBasic IDE. They also              Marathi Corpora Creation and Management
                contain a pseudo-compiled code that is very
                difficult to decipher.                                      We can now read Marathi Dailies i.e.
                      5. Graphics support for drawing lines,          Newspapers (Lokmat, Maharashtra Times,
                circles, boxes and points. These routines are         and Loksatta, Sakal etc) online on the
                the most fun to use. They have a very flexible        Internet. We may need to download fonts if
                syntax. They are almost natural Marathi               required. Unicode addresses this issue
                language instructions. Very easy to read and          regarding the aesthetics of the written textual
                use once you get the hang of them.                    information.
                      6. The find and replace utility searches              We are in process of development of
                only the upper text box which takes in phonetic       Marathi       Sanganakiya Paribhashik
                user input. It does not search in the                 Shabdkosh. Some of the words that we have
                transliterated text box displayed below it. The       found suitable are already put in use in our
                transliteration displayed in the find and replace     communications. We also have found that
                utility is provided for your reference and is not     Janbharati group of CDAC/NCST has
                considered by the MBasic Editor when                  released a Linux version with Marathi icons/
                searching for text.                                   symbols. VPM Thane has also developed and
                                                                      released e-books on works of Vishnushastri
                      7. Sometimes the transliterator can             Chiplunkar. Sanganakiya Marathi Paribhashik
                display garbage in the output window at the           Shabdkosh development is also in progress.
                                                                                                                333
                                                        O gßJUH$ Am{U _amR>rO
                 94